Differentially Expressed Genes in Activin-Induced Apoptotic LNCaP Cells

Differentially Expressed Genes in Activin-InducedA

S*U

R

ibiwaasS(mhstLmc

pdeatishouitadbopa

7

Biochemical and Biophysical Research Communications 257, 187–192 (1999)

Article ID bbrc.1999.0432, available online at http://www.idealibrary.com on

poptotic LNCaP Cells

hilung Lin*,1 and Shao-Yao Ying†Epiclone Inc., 731 S. Chapel Avenue, Suite F, Alhambra, California 91801; and †Department of Cell and Neurobiology,niversity of Southern California, School of Medicine, BMT-401, 1333 San Pablo Street, Los Angeles, California 90033

eceived February 23, 1999

gene transcripts with a small homologous domain, thecbatah

paaittLdttetashetnofirAta

M

DGwTa

Gene transcripts differentially expressed in activin-nduced human prostatic LNCaP apoptotic cells haveeen discovered by an improved subtractive hybrid-zation method, uracil-DNA subtraction assay (USA),hich involves digestion with uracil-DNA glycosylasend mung-bean nuclease. Among the five up-regulatednd seven down-regulated genes, we have identifiedix known (>95% homology and similar size; p16, p53,iva, RHAMM, Pax2, and eIF-4a1), three homologues

>95% homology but different size; myosin, a helicaseotif, and a kinase motif), and three novel genes (noomology). In addition, anti-sense knock-out of a re-ulting novel kinase-like gene was found to abolishhe apoptotic DNA fragmentation in activin-treatedNCaP cells. These findings indicate a new potentialechanism in DNA fragmentation of activin-induced

ell-cycle arresting and apoptosis. © 1999 Academic Press

To this day, detection of genes differentially ex-ressed in the same cell type under two different con-itions has been achieved by methods such as differ-ntial display PCR (1) and representational differencenalysis (RDA) (2, 3). However, these methods areechnically difficult, time consuming and often resultn a great number of identified homologues which areuspected to be differentially expressed. Therefore, weave explored a process to eliminate unwanted homol-gous sequences between compared cDNA libraries bysing non-modified cDNAs as tester and uracil-

ncorporated cDNAs (U-DNAs) as driver. The elimina-ion of non-differentially expressed homologues ischieved by hybridization of the tester to the comparedriver and subsequent digestion of driver-bound hy-rids with uracil-sensitive endonucleases. In this way,nly differentially expressed genes in the tester can bereserved after the enzymatic digestion and final PCRmplification. To prevent cross-over digestion among

1 To whom correspondence should be addressed. Fax: (626) 289-172.

187

ompared cDNA libraries are preferentially restrictedy a four-cutting enzyme (such as Hpa2) and ligated tospecific adaptor to yield sequences ranging from 100

o 700 base pairs with low complexity. This step alsollows for a greater completeness during subtractiveybridization and differential amplification (2).As a model for the USA detection of differentially ex-

ressed genes in apoptosis, we treated LNCaP cells, anndrogen-sensitive human prostate cancer cell line, withctivin. Previously, activin, a member of the transform-ng growth factor b (TGF-b) superfamily, has been showno inhibit cell proliferation and enhance apoptosis inhese cells (4, 5). We have prepared an activin-treatedNCaP cDNA library as tester and an untreated one asriver, and vice versa, in which the driver always con-ained uridine. After subtractive hybridization of testero driver and then enzymatic digestion, the differentiallyxpressed genes were amplified and displayed on an elec-rophoresis gel, from which the results were extractednd confirmed by Northern blot analysis. Based on theequence data of the results from a GenBank search, weave successfully identified twelve highly differentiallyxpressed genes between the activin-treated and un-reated LNCaP cells. However, since the identifiedovel genes may be involved in either cell-cycle arrestr apoptosis, the possible apoptosis-related genes wereurther tested by anti-sense knock-out transfection,n which knocking out a positive apoptotic gene willescue the activin-treated cells from DNA fragmentation.ll information so obtained shed light on the transcrip-

ional control of activin-induced cell growth inhibitionnd apoptosis.

ATERIALS AND METHODS

Oligonucleotides. Four oligonucleotides were used in the uracil-NA subtraction assay as follows: Tester-24mer (59-GCCACCAGAA-AGCGTGTACGTCC-39), Tester-11mer (59-CGGGACGTACA-39ith a dephosphorylated 59-end); Driver-24mer (59-CGGTAGTGAC-CGGTTAAGATCGC-39), Driver-11mer (59-CGGCGATCTTA-39 withdephosphorylated 59-end). The 24mer oligonucleotides were used

0006-291X/99 $30.00Copyright © 1999 by Academic PressAll rights of reproduction in any form reserved.

as 59-adaptors and primers for differential PCR, while the 11mero

tfetFaS(etcthsw0m

fabFncgc(mCf

atpqcpTtdcTiFmpo1a

dfDmamTpc(tbtg

aamHtawocm2d

u

(agndTUps

Vol. 257, No. 1, 1999 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ligonucleotides functioned as linkers for ligation of 59-end adaptors.

General methods. All routine techniques and DNA manipula-ions, including gel electrophoresis, plasmid preparations and trans-ormations, were performed according to standard procedures (6). Allnzymes and buffer treatments were applied following the manufac-ure’s recommendations (Boehringer Mannheim, Indianapolis, IN).or Northern blots, mRNAs were fractionated on 1% formaldehyde-garose gels and transferred onto nylon membranes (Schleicher &chuell, Keene, NH). Probes were labeled with the Prime-It II kit

Stratagene, La Jolla, CA) by random primer extension in the pres-nce of [32P]-dATP (.3000 Ci/mM, Amersham International, Arling-on Heights, IL), and purified with Micro Bio-Spin chromatographyolumns (Bio-Rad, Hercules, CA). Hybridization was carried out inhe mixture of 50% freshly deionized formamide (pH 7.0), 5 3 Den-ardt’s solution, 0.5% SDS, 4 3 SSPE and 250 mg/ml denaturedalmon sperm DNAs (18 h, 42°C). Membranes were sequentiallyashed twice in 2 3 SSC, 0.1% SDS (15 min, 25°C), and once each in.2 3 SSC, 0.1% SDS (15 min, 25°C); and 0.2 3 SSC, 0.1% SDS (30in, 65°C) before autoradiography.

Cell culture and activin treatment. LNCaP cells were obtainedrom the American Type Culture Collection (ATCC, Rockville, MD)nd grown in RPMI 1640 medium supplemented with 10% fetalovine serum with 100 mg/ml gentamycin at 37°C under 10% CO2.or three-day activin induction, LNCaP cells were treated with 200g/ml activin per day, while other cells were treated with medium asontrol. On the fifth day after the first treatment, a 56% reduction inrowth was observed in the activin-treated cells compared to theontrol by both microscopy and cell counting as previously reported4). The two groups of cells were independently trypsinized andRNAs were purified by poly-(dT) dextran columns (Qiagen, Santalarita, CA). The quality of isolated mRNAs was assessed on 1%

ormaldehyde-agarose gels.

Generation of double-stranded cDNA libraries and representativemplicons. The first strand of cDNAs was prepared by reverseranscription of the mRNAs with oligo(dT) primers following therotocol of a cDNA Cycle kit (Invitrogen, Carlsbad, CA), and itsuality was assessed on a 2% agarose gel. The second strand ofDNAs was synthesized with an enzyme cocktail containing DNAolymerase I, RNase H and T4 ligase, as reported by Gubler et al. (7).o generate adequate lengths of cDNA amplicons for efficient sub-raction and amplification, double-stranded cDNAs (1.5 mg) wereigested with Hpa2 (3 hr, 37°C), recovered by 100 bp-cutoff microcon-entrator columns (Amicon, Beverly, MA) and then ligated to eitherester-24/11mer or Driver-24/11mer adaptors in a mixture contain-

ng the 24/11mer oligo (0.75 nmol each) in 30 ml of 1 3 ligase buffer.or precise ligation between restricted cDNAs and the adaptors, theixture was held by gradually cooling from 50°C to 10°C over a

eriod of one hour, and then T4 ligase was added to anneal the 24merligonucleotides onto the 59-ends of the restricted cDNAs, at 16°C for4 hr. This formed the representative amplicons for both the testernd driver respectively, depending on the distinctive adaptor used.

Generation of uracil-cDNA driver. For incorporation of deoxyuri-ylate into the driver cDNAs, multiple PCR reactions were set up asollows: each 50 ml reaction containing driver amplicon (10 ng), 1 mMriver-24mer oligo, dNTP mixture 1 (0.2 mM dATP, 0.2 mM dGTP, 0.2M dCTP, 0.05 mM dTTP and 0.5 mM dUTP) and Taq DNA polymer-

se (5 U) in 1 3 PCR buffer (Fig. 1). The Driver-11mer linker waselted away (3 min, 72°C), and the recessed 39-ends were filled in withaq DNA polymerase (3 min, 72°C). Thirty cycles of amplification wereerformed (1 min, 95°C; 3 min, 72°C), and the PCR products wereombined and recovered by a microconcentrator column in Tris buffer10mM, pH 7.0). The driver-adaptor was removed by Hpa2 cleavage andhe digest was recovered by a microconcentrator column in EE 3 3uffer (30 mM EPPS, pH 8.0 at 20°C; 3 mM EDTA) at 1 mg/ml to formhe driver. The quality of the driver (2 mg) was assessed on a 2% agaroseel, ranging from 100 bp to about 1 kb.

188

Subtractive hybridization, enzymatic digestion and differentialmplification. For the first subtractive hybridization, 0.2 mg of testermplicon (tester) was mixed with 10 ml restricted driver, overlaid withineral oil, denatured (5 min, 98°C) and immediately cooled on ice.ybridization was performed with the phenol emulsion reassociation

echnique in a 400 ml solution containing 1.5 M NaSCN and 8% phenolt 25°C for 48 hr (8, 9). An emulsion of the phenol and aqueous phasesas maintained throughout the hybridization by continuous agitationn a vortex mixer. The hybridized DNAs were recovered by a microcon-entrator column in 20 ml of Tris buffer and treated with DNA poly-erase I-T4 DNA polymerase 3:1 mixture (3 min, 37°C without dNTPs;

5 min, 37°C with dNTP mixture 2 in 0.2 mM each for dATP, dGTP,CTP and dTTP) to fill in the 39-end of the tester.In order to eliminate driver homologues in the hybridized DNAs,

racil-DNA glycosylase (UNG) was added (30 min, 25°C) to remove

FIG. 1. Schematic protocol for uracil-DNA subtraction assayUSA), illustrating the sequential enzymatic digestion and differentialmplification steps after subtractive hybridization. The enzymatic di-estion contains two substeps: uracil-DNA glycosylase digestion whichicks all driver-like homologues, and single-strand-specific nucleaseigestion which cleaves the homologues into unamplifiable fragments.he process is outlined through the final products of the first round ofSA. To reiterate another round of subtraction, the first differenceroducts are used as tester following the same scheme to generate theecond difference products and so on.

uracil from driver-driver and driver-tester hybrid duplexes, resultingigtncpddbiTespapw

UIpsdsp

fipaDtapbrcqtsmtDMm

R

I

llcotaammwt

dcotlm((uvarg(

aiaiiaw(tig

Fualattpaddodfms


n a partially single-stranded conformation with abasic nicks andaps. The abasic homologue duplexes were then subjected to diges-ion by a single-strand-specific endonuclease, such as mung-beanuclease (MBN) and S1 nuclease (2 U at 25°C for 20 min), and finallyleaved into unamplifiable fragments (,50 bp). To ensure the com-leteness of homologue elimination, the hybridization and enzymaticigestion described above were repeated at least once. The finaligest was recovered by a microconcentrator column in 20 ml of Trisuffer and prepared for differential PCR in a 50 ml reaction, contain-ng 2 ml of the digest, 1 mM Tester-24mer oligo, dNTP mixture 2 andaq DNA polymerase. Final PCR products were phenol-extracted,thanol-precipitated, resuspended in 20 ml of Tris buffer and as-essed on a 3% agarose gel. The DNA bands shown on the electro-horesis gel were excised, recovered by a gel extraction kit (Qiagen)nd further purified by a 4% non-denaturing polyacrylamide gel. Therocesses of hybridization, enzymatic digestion and amplificationere repeated until a clear banding pattern was observed.

Cloning and sequencing of difference products. Final products ofSA were ligated to the pCR2.1 plasmid and transformed into

NVaF9 cells using a TA cloning kit (Invitrogen). Double-strandedlasmid DNAs were purified by miniprep spin columns (Qiagen), andequenced by a Sequenase v.2 DNA sequencing kit (Amersham) withideoxy-mediated chain termination. Resulting sequences wereearched and compared to the Genbank database using the BLASTrogram from the National Institutes of Health (NIH).

Anti-sense knock-out assay. Following the single-stranded ampli-cation method reported by Medori et al. (10), 100 ng of each USAroduct was used in a 100 ml PCR reaction containing 30 pmol ofnti-sense primer, 0.3 pmol of sense primer, dNTP mixture 2, TaqNA polymerase and 1.5 mM MgCl2. A thirty cycle PCR amplifica-

ion was carried through denaturation at 95°C, annealing at 55°Cnd extension at 72°C for 1 min in each step. The anti-sense PCRroduct was recovered by a microconcentrator column in Hepesuffer (20 mM, pH 7.4), and modified with a covalent modificationeagent (Epiclone, Alhambra, CA) to introduce covalent bondingapability between the modified probe and its target nucleotide se-uence. Transfection was carried out with the DOTAP liposomalransfection reagent (Boehringer Mannheim) using cationic lipo-ome-mediated intracellular transport of the anti-sense probe (0.7g/ml) into activin-treated LNCaP cells. After three-days of activinreatment and two-days of incubation as described above, genomicNAs were isolated by an apoptotic DNA ladder kit (Boehringerannheim) and assessed on a 2% agarose gel. The cell growth andorphology were also examined.

ESULTS AND DISCUSSION

dentification of Differentially Expressed Transcriptsbetween Activin-Treated and Non-Treated LNCaPCells by USA

As shown in Fig. 2, when an activin-treated cDNAibrary was used as the tester and an untreated cDNAibrary was used as the driver, three bands whichontained thirteen gene fragments (Fig. 2, lane i) werebtained only in the activin-treated cells, including fiveranscriptional products whose levels were increasedt least two fold above the controls. These products areCCPK (cell-cycle check-point protein kinase)-like ho-ologue (LT6), p16 (LT7), p53 (LT11), a Siva-like ho-ologue (LT14) and a novel gene. On the other hand,hen an untreated cDNA library was used as the

ester and an activin-treated cDNA library as the

189

river, one thick band and two weaker bands whichontain fifteen gene fragments (Fig. 2, lane h) werebtained only in the untreated cells, including sevenranscriptional products which were decreased to asow as 45% of the controls. These products are a

yosin-like homologue (LC2), hyaluronan receptorLC3), a helicase motif-like homologue (LC9), Pax2LC12), eIF-4A1 (LC13) and two novel genes. The fivep-regulated gene products represented the genes in-olved in enhancing activin-induced cell-cycle arrestnd apoptosis (11, 12, 13, 14), whereas the seven down-egulated genes indicated that the expression of theseenes would be reduced by cell cycle arrest or apoptosis15, 16, 17, 18, 19).

Previously, p53 and p16 have been shown by us (13)nd others (12) to be up-regulated in activin-mediatednhibition of cell proliferation and enhancement ofpoptosis in LNCaP cells. Indeed, both genes weredentified by the USA method and are known to benvolved in the G1 phase arrest of the cell cycle as wells the induction of apoptosis. No sequence was foundhen a cDNA library was subtracted with itself

treated to treated and untreated to untreated, respec-ively) after thirty cycles of PCR amplification, suggest-ng that the identification of differentially expressedenes is highly specific. In addition, the cross-reaction be-

FIG. 2. Agarose gel electrophoresis of PCR subtracted products.rom left to right: lane a, DNA markers; lane b, driver amplicon fromntreated LNCaP cells amplified by driver-24 primer; lane c, testermplicon from activin-treated cells amplified by tester-24 primer;ane d, driver amplicon amplified by tester primer; lane e, testermplicon amplified by driver primer; lane f, driver amplicon sub-racted with driver and then amplified by driver primer; lane g,ester amplicon subtracted with tester and then amplified by testerrimer; lane h, driver amplicon subtracted with tester and thenmplified by driver primer; lane i, tester amplicon subtracted withriver and then amplified by tester primer. The self-subtraction ofriver to driver (f) and tester to tester (g) shows complete eliminationf all sequences, while the mutual subtraction between tester andriver (h, i) presents difference products on the gel, indicating dif-erential gene expressed in the tester and driver respectively. Theisuse of PCR primer (d, e) will result in no amplification due to the

pecific affinity of the primer to its own adaptor.

tsatsstlttn

tbaaiglahp

ssitai


ween tester and driver was completely inhibited bypecially designed adapters and primers which werebsolutely inert to each other under our PCR condi-ions (Fig. 2, lanes d & e). Moreover, there was noignificant difference between the first round and theecond round of USA in the present model, indicatinghat a complete subtraction was achieved. Neverthe-ess, when cDNA libraries derived from tissues, ratherhan cells, were used, more than one round of sub-raction was required to obtain clear-cut results (dataot shown).

FIG. 3. Autoradiogram of positive Northern blots hybridized to theven down-regulated genes mainly present in untreated LNCaP cellshows five up-regulated genes significantly increased after activin tren the activin-treated LNCaP cells. The down-regulated known genehe up-regulated known genes (LT6, 7, 11, 14) are involved in either cltered above two folds. The size of each identified gene transcript isndicates the sequence similarity between the identified fragment a

190

CCPK, a serine-threonine kinase involved in regula-ion of the cell cycle, plays a role in genetic interactionetween the regulation of mitosis, cell differentiation,nd apoptosis (11, 21, 22). Siva is a proapoptotic proteinnd overexpression of this molecule resulted in apoptosisn various cell lines (14). A myosin regulatory light chainene has been isolated from a normal human prostateibrary (23) which has been reported to block the TNFctivation of DNA fragmentation (24). The receptor foryaluronan-mediated motility (RHAMM) has been re-orted to mediate migration, transformation, and meta-

nal difference products of USA. Upper panel (LC2 to LC13) indicatest not in the activin-treated cells, while the lower panel (LT1 to LT14)ent. p16 (LT7) and p53 (LT11) have been known to be up-regulatedC2, 3, 9, 12, 13) are related to physiological functions of cells, and

cycle regulation, apoptosis or both. Genes listed are transcriptionallyuced from individual Northern blots, and the homology shown here

its deduced gene, rather than the full identified sequence.

e fibuatms (L

ell-dednd

static spread of murine fibroblasts (16) and is expressediftpgti

caatoiiivttth

C

blibfwftswiravmNw

F

pshiLhw

estwtofpL

satsgNceidogdciccp

ocLttsTfaL


n breast cancer cells (25). Helicase motif is a gene thatunctions in the maintenance of genome stability and inhe suppression of illegitimate recombination. The ex-ression of Pax2 gene has been reported to reduce cellrowth of renal carcinoma cells (18, 19). Eukaryoticranslation initiation factor 4A1 is thought to be involvedn tumor development (19, 26, 27).

As observed previously, the G1 phase arrest of theell cycle occurs in the third day after activin treatmentnd apoptosis begins early in the fifth day. Presum-bly, the identified genes in our model are related tohe G1 phase arrest and also the late activation phasef apoptosis. Our results demonstrate that most of thedentified up-regulated genes are likely to be involvedn either G1 arrest, apoptosis or both. We also havedentified several novel genes which may also be in-olved in cell cycle arrest. Indeed, the identification ofhese genes is consistent with our previous observa-ions that activin affects cell growth in LNCaP cellshrough the inhibition of cell proliferation and en-ancement of apoptosis.

onfirmation of Differential Expression Levelsby Northern Blot Hybridization

As shown in Fig. 3, all identified genes were confirmedy Northern blot hybridization (N 5 3 or 4) onto mRNAibraries from activin-treated and untreated cells, provid-ng direct evidence for the alterations of gene expressionetween activin-treated and untreated LNCaP cells. Aragment showing at least two-fold changes (P , 0.01)as considered positive. Thirteen clones were identified

rom activin-treated cells and fifteen clones from non-reated control cells. Of which twelve positive cloneshowed 2-8 fold alterations of mRNA expression, amonghich five were up-regulated and seven down-regulated

n activin-induced cells. As listed in Fig. 3, the five up-egulated genes are related to either cell-cycle arrest orpoptosis, and the seven down-regulated genes are in-olved in the maintenance of normal cell physiology. TheRNA length of a positive clone was deduced from theorthern blots and then used to match its original geneith sequence data from Genbank.

unctional Assay of Up-Regulated Novel Genesfrom Apoptotic Cells by Anti-SenseKnock-Out Transfections

Since the difference products identified by USA areartial fragments restricted from the original gene tran-cripts, by using anti-sense knock-out transfection weave tested the functions of the difference products

n preventing the apoptotic effects on activin-treatedNCaP cells. From USA results, clone LT6 (a CCPK-likeomologue) was expressed only in apoptotic LNCaP cellshile clone LC9 (a helicase-motif-like homologue) was

191

xpressed only in the proliferative LNCaP cells. Ashown in Fig. 4, the anti-sense of LT6 rescued activin-reated LNCaP cells from apoptotic DNA fragmentation,hereas that of LC9 further increased the fragmenta-

ion. The sense control of both genes showed no influencen apoptotic DNA fragmentation, indicating that the ef-ects of knock-out are sequence-specific. For operationalurposes, clone LT6 was referred to as apoptosin whereasC9 apoptostatin.It is noteworthy that the results presented demon-

trate a totally new approach in our understanding ofpoptosis as wells as genes in cancer development andreatment. The full-length mRNA sequence of apopto-in and apoptostatin as well as the functions of theseenes in normal and cancer cells remain to be defined.evertheless, the USA method provides a more effi-

ient and sensitive means for identifying differentiallyxpressed genes. Potentially, this method can be usedn identifying different gene expressions involved inevelopment, cell differentiation, aging, and a varietyf pathological disorders, such as cancer formation,enetic defects, autoimmune diseases, and any otherisorders related to genetic malfunction. The identifi-ation of these differentially expressed genes will helpn the determination of their open-reading frames andorresponding peptides which may contribute to a spe-ific drug-design or therapy for regulation of the ex-ression of these genes.

FIG. 4. Distinctive functions detected after anti-sense knock-outf LT6 and LC9. The antisense of LT6 rescues activin-treated LNCaPells from apoptotic DNA fragmentation (lane D), whereas that ofC9 increases the fragmentation (lane F). Both transfection of con-rol and sense sequences shows no effect on apoptotic DNA fragmen-ation (lane C, activin-treated positive control; lane E, result ofense-LT6 transfection; lane G, result of sense-LC9 transfection).he result of antisense-LT6 transfection is similar to normal genome

rom untreated cells (lane B), indicating a significant inhibition ofpoptotic DNA fragmentation after blocking the gene transcript ofT6 in activin-treated LNCaP cells.

REFERENCES

1

1

1

1

1

S., and Schlossman, S. F. (1997) Proc. Natl. Acad. Sci. USA 94,

1

1

1

1

1

2

22

22

2

2

2


1. Liang, P., and Pardee, A. B. (1992) Science 257, 967–971.2. Lisitsyn, N., Lisitsyn, N., and Wigler, M. (1993) Science 259,

946–951.3. Hubank, M., and Schatz, D. G. (1994) Nucleic Acids Research 22,

5640–5648.4. Zhang, Z., Zheng, J., Zhao, Y., Li, G., Batres, Y., Luo, M. P., Wan,

M., and Ying, S. Y. (1997) Int. J. of Oncol. 11, 727–736.5. Wang, Q. F., Tilly, J. L., Preffer, F., Schneyer, A. L., Crowley,

W. F., Jr., and Sluss, P. M. (1996) Endocrinology 137, 5476–5483.

6. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) MolecularCloning, a laboratory manual, Cold Spring Harbor LaboratoryPress, Cold Spring Harbor.

7. Gubler, U., and Hoffman, B. J. (1983) Gene 25, 263–269.8. Kohne, D. E., Levison, S. A., and Byers, M. J. (1977) Biochemis-

try 16, 5329–5341.9. Travis, G. H., and Sutcliffe, J. G. (1988) Proc. Natl. Acad. Sci.

USA 85, 1696–1700.0. Medori, R., Tritschler, H. J., and Gambetti, P. (1992) BioTech-

niques 12,132–135.1. Lewis, B. C., Shim, H., Li, Q., Wu, C. S., Lee, L. A., Maity, A., and

Dang, C. V. (1997) Mol. Cell. Biol. 17, 4967–4978.2. Itoh, N., Kakehi, Y., Akao, T., Kinoshita, H., Okada, Y., and

Yoshida, O. (1997) Jpn. J. Cancer Res. 88, 229–233.3. Ying, S. Y., Zhang, Z., Batres, Y., Zhao, Y., Lin, S. L., and Li, G.

(1997) Int. J. Oncol. 11, 591–595.4. Prasad, K. V. S., Ao, Z., Yoon, Y., Wu, M. X., Rizk, M., Jacquot,

192

6346–6351.5. Wang, A., Liang, Y., Fridell, R., Probst, F. J., Wilcox, E. R.,

Touchman, J. W., Morton, C. C., Morell, R. J., Noben-Trauth, K.,Camper, S. A., and Friedman, T. B. (1998) Science 280, 1447–1451.

6. Wang, C., Entwistle, J., Hou, G., Li, Q., and Turley, E. A. (1996)Gene 174, 299–306.

7. Sanchez-Alonso, P., and Guzman, P. (1998) Genetics 148, 1043–1054.

8. Eccles, M. R., Wallis, L. J., Fidler, A. E., Spurr, N. K., Goodfel-low, P. J., and Reeve, A. E. (1992) Cell Growth Differ. 3, 279–289.

9. Gnarra, J. R., and Dressler, G. R. (1995) Cancer Res. 55, 4092–4098.

0. Greaves, D. R., Quinn, C. M., Seldin, M. F., and Gordon, S.(1998) Genomics 15, 165–168.

1. Efimov, V. P., and Morris, N. R. (1998) Genetics 149, 101–116.2. Donadelli, R., Benatti, L., Remuzzi, A., Morigi, M., Gullans,

S. R., Benigni, A., Remuzzi, G., and Noris, M. (1998) Biochem.Biophys. Res. Commun. 246, 881–887.

3. NIC-CGAP Tumor Gene Index. (1997) g2583404. [Unpublished]4. Wright, S. C., Zheng, H., Zhong, J., Torti, F. M., and Larrick,

J. W. (1993) J. Cell. Biochem. 53, 222–233.5. Assmann, V., Marshall, J. F., Fieber, C., Hofmann, M., and Hart,

I. R. (1998) Cell. Sci. 111, 1685–1694.6. Jones, E., Quinn, C. M., Montgomery, D. S., Ford, M. J., Kollble,

K., Gordon, S., and Greaves, D. R. (1998) Genomics 53, 248–250.7. Eberle, J., Krasagakis, K., and Orfanos, C. E. (1977) Int. J.

Cancer 71, 396–401.

Documents

Differentially Expressed Genes in Activin-Induced Apoptotic LNCaP Cells